A method and device form a prediction value. A prediction direction is locally described by nonlinear trajectories. The prediction direction can be used in forming a prediction value to achieve a more accurate prediction determination. The method and device can be used in image compression or image-sequence compression.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method for forming a prediction value for a first image value of a first pixel of an image on the basis of second image values of respective reconstructed pixels of the image, the method comprising: forming at least two trajectories, each trajectory having a third image value for points along the trajectory, each trajectory having a shape that is not a straight line; assigning a second image value to each third image value and thereby each trajectory; selecting one of the trajectories as an auxiliary trajectory, the auxiliary trajectory being the trajectory at a shortest distance from a position of the first pixel; and generating the prediction value based on the third image value of the auxiliary trajectory.
2. The method as claimed in claim 1 , wherein for assigning the second image value, that second image value is selected either which minimizes a distance between the second image value and the trajectory assigned to the third image value or which corresponds to a reconstructed pixel of the first pixel.
3. The method as claimed in claim 1 , wherein only one auxiliary trajectory is taken into account for generating the prediction value if the distance from the position of the first pixel is equal to zero, and if the distance from the position of the first pixel is equal to zero, the third image value of the auxiliary trajectory is assigned to the first image value.
4. The method as claimed in claim 1 , wherein at least n plural auxiliary trajectories are selected, the n plural auxiliary trajectories being the trajectories at n shortest distances from the position of the first pixel, and the prediction value is generated based on the third image values of the n plural auxiliary trajectories.
5. The method as claimed in claim 4 , wherein the prediction value is generated using a weighted averaging of the third image values of the n plural auxiliary trajectories.
6. The method as claimed in claim 5 , wherein the weighted averaging uses reciprocal values of the respective distances of the auxiliary trajectories from the position of the first pixel.
7. The method as claimed in claim 1 , wherein a new trajectory is formed in such a way that: the first pixel is a point on the new trajectory, and the third value of the new trajectory is equivalent to the prediction value or equivalent to an image value for a reconstructed first pixel.
8. The method as claimed in claim 7 , wherein two auxiliary trajectories are selected, the two auxiliary trajectories being the trajectories at two shortest distances from the position of the first pixel, and the new trajectory is arranged between the two auxiliary trajectories in such a way that at an arbitrary point on the new trajectory, a ratio of the shortest distances between the arbitrary point and the two auxiliary trajectories is identical to a ratio of the distances of the two auxiliary trajectories from the position of the first pixel.
9. The method as claimed in claim 1 , wherein each second image value is the image value assigned to the reconstructed pixel from which the trajectory extends.
10. A method for forming a prediction value for a first image value of a first pixel of an image on the basis of a second image value of a reconstructed pixel of the image, the method comprising: forming at least two trajectories each trajectory having a third image value for points along the trajectory, each trajectory having a shape that is not a straight line; selecting one of the trajectories as an auxiliary trajectory, the auxiliary trajectory being the trajectory at a shortest distance from a position of the first pixel; assigning the second image value to the third image value of the auxiliary trajectory; and generating the prediction value based on the third image value of the auxiliary trajectory.
11. A device to form a prediction value for a first image value of a first pixel of an image on the basis of second image values of respective reconstructed pixels of the image, the device comprising: a first unit to form at least two trajectories, each trajectory having a third image value for points along the trajectory, each trajectory having a shape that is not a straight line; a second unit to assign a second image value to each third image value and thereby each trajectory; a third unit to select one of the trajectories as an auxiliary trajectory, the auxiliary trajectory being the trajectory at a shortest distance from a position of the first pixel; and a fourth unit to generate the prediction value based on the third image value of the auxiliary trajectory.
12. The device as claimed in claim 11 , wherein at least n plural auxiliary trajectories are selected, the n plural auxiliary trajectories being the trajectories at n shortest distances from the position of the first pixel, and the prediction value is generated based on the third image values of the n plural auxiliary trajectories.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
January 5, 2012
December 29, 2015
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.